It has a table of attributes and a listing immediately following
class quantecon.markov.core.MarkovChain(P)[source]
Bases: object
Class for a finite-state discrete-time Markov chain. It stores useful information such as the stationary distributions, and communication, recurrent, and cyclic classes, and allows simulation of state transitions.
Parameters:
P : array_like or scipy sparse matrix (float, ndim=2)
The transition matrix. Must be of shape n x n.
Notes
In computing stationary distributions, if the input matrix is a sparse matrix, internally it is converted to a dense matrix.
Attributes
P (ndarray or scipy.sparse.csr_matrix (float, ndim=2)) See Parameters
stationary_distributions (array_like(float, ndim=2)) Array containing stationary distributions, one for each recurrent class, as rows.
is_irreducible (bool) Indicate whether the Markov chain is irreducible.
num_communication_classes (int) The number of the communication classes.
communication_classes (list(ndarray(int))) List of numpy arrays containing the communication classes.
num_recurrent_classes (int) The number of the recurrent classes.
recurrent_classes (list(ndarray(int))) List of numpy arrays containing the recurrent classes.
is_aperiodic (bool) Indicate whether the Markov chain is aperiodic.
period (int) The period of the Markov chain.
cyclic_classes (list(ndarray(int))) List of numpy arrays containing the cyclic classes. Defined only when the Markov chain is irreducible.
Methods
simulate(ts_length[, init, num_reps, ...]) Simulate time series of state transitions.
cdfs
cdfs1d
communication_classes
cyclic_classes
digraph
is_aperiodic
is_irreducible
num_communication_classes
num_recurrent_classes
period
recurrent_classes
Upon closer inspection it looks like any method that doesn't include a docstring is not listed in the method table that is constructed. And the ones that do have docstrings are in twice.
Example: quantecon.markov.core.MarkovChain
It has a table of attributes and a listing immediately following